Listener stops responding to new links after some time about rnsh HOT 5 CLOSED

After further testing, I don't think this is a problem with the listener. I think the application is actually timing out waiting for a path. A loop to try a few times might be needed. It would probably be helpful to add back in the spinner and status messages from the rnx code to help clue the user in.

What I've noticed is that if it fails to connect, I can try a couple more times and it eventually does connect. 🤔

from rnsh.

Hmm, that is strange... I will try looking a bit into this as well.

from rnsh.

I have a little more information on this. It seems to be related to the number of hops involved. When I set up a direct connection to the RNS instance on the BBS server, I no longer had connection issues.

from rnsh.

Hmm, I wonder if this could be caused by the path running over the I2P links initially, and then something causing those to become unavailable, after which the connection fails (probably until a new announce is sent, re-establishing the path over the TCP links).

I have seen weird flapping going on with the I2P interfaces at times, but haven't been able to narrow it down to anything concrete yet.

from rnsh.

After a lot of testing, this does indeed appear to be due to delays in path discovery. It's just part of running on Reticulum.

from rnsh.